Abstract:
In this paper, we present a digital hardware implementation of an artificial neuron on-chip unsupervised trained with Hebbian rule. The main characteristics of this solution are on-chip learning algorithm implementation and high reconfiguration capability and operation under real time constraints.